Electrode Mass Balancing as an Inexpensiveand Simple Method to Increase theCapacitance of Electric Double-LayerCapacitorsBritta Andres1*, Ann-Christine Engstrom1, Nicklas Blomquist1,2, Sven Forsberg1,
Christina Dahlstrom3, Håkan Olin1
1 Department of Natural Sciences, Mid Sweden University, Sundsvall, Sweden, 2 STT Emtec AB, Sundsvall,
Sweden, 3 Department of Chemical Engineering, Mid Sweden University, Sundsvall, Sweden
AbstractSymmetric electric double-layer capacitors (EDLCs) have equal masses of the same active
material in both electrodes. However, having equal electrode masses may prevent the
EDLC to have the largest possible specific capacitance if the sizes of the hydrated anions
and cations in the electrolyte differ because the electrodes and the electrolyte may not be
completely utilized. Here we demonstrate how this issue can be resolved by mass balanc-
ing. If the electrode masses are adjusted according to the size of the ions, one can easily
increase an EDLC’s specific capacitance. To that end, we performed galvanostatic cycling
to measure the capacitances of symmetric EDLCs with different electrode mass ratios
using four aqueous electrolytes— Na2SO4, H2SO4, NaOH, and KOH (all with a concentra-
tion of 1 M)—and compared these to the theoretical optimal electrode mass ratio that we
calculated using the sizes of the hydrated ions. Both the theoretical and experimental val-
ues revealed lower-than-1 optimal electrode ratios for all electrolytes except KOH. The larg-
est increase in capacitance was obtained for EDLCs with NaOH as electrolyte. Specifically,
we demonstrate an increase of the specific capacitance by 8.6% by adjusting the electrode
mass ratio from 1 to 0.86. Our findings demonstrate that electrode mass balancing is a sim-
ple and inexpensive method to increase the capacitance of EDLCs. Furthermore, our
results imply that one can reduce the amount of unused material in EDLCs and thus
decrease their weight, volume and cost.
Introduction
Different methods to enhance the performance of supercapacitors have been recently reported.New advanced and highly porous electrodematerials [1, 2] have been developed, and functio-nalized electrode surfaces have been employed in supercapacitors [3]. The importance ofmatching electrolytes and electrodematerials has been discussed, and Chmiola et al. [2, 4, 5]
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 1 / 12
and Largeot et al. [6] have reported the importance of matching the electrode pore size to thesize of the electrolyte ions or vice versa. These adjustments result in enhanced device perfor-mance. However, one must consider that most high-performance supercapacitors containexpensive and toxic materials. Thus, simple methods to improve a supercapacitor’s perfor-mance are desired when the choice of materials is limited due to factors such as cost savings,the environmental impact of the materials, or their availability.
In this study, we present a simple and effective approach to increase the capacitance of elec-tric double-layer capacitors (EDLCs). Depending on the applied electrolyte, one should adjustthe weight of the electrodes according to the size of the electrolyte ions to fully utilize the sur-face of both electrodes. The mass balancing will result in an increased capacitance.
Perspective and purpose of this study
The main objective of this study is to produce and optimize inexpensive and environmentallyfriendly EDLCs for the automotive industry. One of today’s largest challenges of the automo-tive industry is the transition from the traditional internal combustion engine using fossil fuelsto electric drives. Several factors still hinder this development. In addition to limited drivingranges and the poor infrastructure of charging stations in many regions, the high price ofenergy storage devices limits the commercialization of electric vehicles. Different types ofenergy storage devices, such as batteries and fuel cells, have been tested in electric vehicles.Supercapacitors are often proposed as intermediate power storage in combination with batter-ies, fuel cells or other long-term energy storage systems. Due to their high power density,supercapacitors can complement and protect electrochemical energy storage devices.
The EDLCs designed in this study are developed to be used as temporary energy storagedevices in kinetic energy recovery systems (KERSs) in cars. Via a generator, the braking energyof a car can be converted into electrical energy, which can be stored in an EDLC. The storedenergy can be used to accelerate the car using an electricmotor or to power its on-board sys-tems. Due to strict manufacturer demands on the weight, volume and price of supercapacitors,we focused on inexpensivematerials, which are sufficient for this application. Althoughmateri-als that offer a better performance in EDLCs are available, they do not meet the demandsreviewed above. Furthermore, EDLCs should be maintenance free, reliable and highly efficient.In addition to these criteria, we aim to produce environmentally friendly and recyclableEDLCs.
Mass-balancing principle
Symmetric EDLCs are composed of electrodes that contain the same active material in boththe positive electrode and the negative electrode and use equal masses of active material inboth electrodes.However, this electrodemass ratio might not be the optimal choice unless theelectrolyte anions and cations have the same size. If the size of the electrolyte ions differs, oneof the electrodesmight not be fully covered with ions; thus, it may not be completely utilized.By contrast, there might be an excess of ions that do not contribute to the overall capacitanceof an EDLC. In the case of smaller ions, the oppositely charged electrode is not completely cov-ered with the smaller ions, whereas the other electrode is fully covered with the larger ions. Ifboth electrodes are fully covered, there will be an excess of larger ions. These ions will not con-tribute to the capacitance.
To effectively use the entire surface area of both electrodes and to avoid excess electrolyte,the electrodemass ratio should be adjusted to the ion size ratio. The performance optimizationof supercapacitors by mass balancing is widely used in asymmetric supercapacitors, e.g., hybridsupercapacitors, in which different active materials are used for the positive and negative
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 2 / 12
(http://www.lansstyrelsen.se/vasternorrland/
En/Pages/default.aspx), Bo Rydin Foundation
(http://www.sca.com/en/Career/Researcher/
Bo_Rydin_foundation/), STT Emtec (http://
www.sttemtec.com/en/home/home.php),
Superior Graphite (http://www.
superiorgraphite.com/index.html), Nordic
Paper (http://www.nordicpaper.com/) and SCA
(http://www.sca.com/en/Home/). The authors
BA, ACE, NB and SF are funded in the KEPS
project. The funders had no role in study
design, data collection and analysis, decision to
publish, or preparation of the manuscript. The
funder STT Emtec provided support in the form
of salaries for authors [NB], but did not have
any additional role in the study design, data
collection and analysis, decision to publish, or
preparation of the manuscript. The specific
roles of this author is articulated in the ‘author
contributions’ section.
Competing Interests: The project, which this work
is included in, has four industrial co-funders (STT
Emtec AB, Superior Graphite, Nordic Paper, and
SCA). This does not alter the authors’ adherence to
electrodes [7–10]. Due to the differing electrode properties, such as capacitance and potentialrange, the electrodemasses need to be balanced. Electrodemass balancing is employed towiden the voltage window of hybrid supercapacitors and pseudocapacitors, although it canalso be applied for EDLCs [7–13]. Furthermore, a larger voltage window results in an increasedenergy density.
Experimental setup and choice of materials
We investigated the optimal electrodemass ratio in different aqueous electrolytes. Aqueouselectrolytes are an attractive alternative to organic electrolytes that are commonly used insupercapacitors. Most commercial supercapacitors use expensive and toxic electrolytes, such astetraethylammonium tetrafluoroborate (TEA BF4) in acetonitrile [14]. However, only aqueouselectrolytes qualify for application in inexpensive and environmentally friendly EDLCs. Theyoffer many advantages, such as good ionic conductivity, low toxicity, non-flammability andlow cost. Furthermore, supercapacitors with aqueous electrolytes offer higher capacitances andhigher power densities than organic electrolytes. The following electrolytes were tested: sodiumsulfate (Na2SO4), sulfuric acid (H2SO4), sodium hydroxide (NaOH) and potassium hydroxide(KOH). Potassium hydroxide and sulfuric acid were chosen because they are the most commonaqueous electrolytes used in EDLCs [14]. Sodium sulfate was tested because it has an almostneutral pH, which is favorable for some electrodematerials. Sodiumhydroxide was investi-gated to complement the test results of sodium sulfate and potassium hydroxide. By choosinganother anion or cation instead of the sulfate ion and the potassium ion, respectively, weexpected to observe the influence of the ions.
The electrodematerial was a mixture of graphite and activated carbon with cellulose nanofi-bers (CNFs), also termed cellulose nanofibrils or nanofibrillated cellulose, as the binder. CNFswere used as the binder because they are an environmentally friendly alternative to fluorinatedpolymers, such as polytetrafluoroethylene (PTFE), that are commonly used in supercapacitors[14, 15]. Moreover, CNFs have previously demonstrated goodmechanical and electrical prop-erties as a binder in graphite electrodes [16]. An initial study showed that the addition of 10%CNFs, in relation to the amount of active material, exhibited the best performance in the testedcomposite. Graphite and activated carbon were chosen as the active material for the electrodes.Initial tests of these materials showed that a composite with 50% graphite and 50% activatedcarbon achieved both a good electrical conductivity and a favorable capacitance. Usually, themain component in EDLC electrodes is activated carbon with a high surface area, which pro-vides a high capacitance. Increasing the amount of activated carbon, however, results in areduced electric conductivity and a large resistance. Previously, we reported that the graphiteused in this study exhibits a low sheet resistance and low electrical resistivity [16, 17]. Theseconditions are favorable for achieving a high power density. For the application of EDLCs inKERSs, a high power density is preferred over a high capacitance. Furthermore, EDLCs con-structed from this graphite exhibit a good capacitance, cyclability and efficiency [16]. Thus,composites with equal amounts of graphite and activated carbon and an additional 10% CNFswere used for the electrodes in this study.
Supercapacitors with different electrodemass ratios were prepared, and galvanostaticcycling was performed to evaluate its influence on the capacitance of the EDLCs. Here, the elec-trodemass ratio is defined as the mass of the positive electrode divided by the mass of the nega-tive electrode; see Eq (3). Galvanostatic cycling was chosen as the test procedure because it is afast and simple method to determine the capacitance of supercapacitors. A two-electrode setupwas used to imitate realistic operating conditions. By contrast to three-electrodemethods, suchas cyclic voltammetry, two-electrode galvanostatic cycling measures the entire cell instead of
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 3 / 12
the performance of a single electrode. This approach is advantageous because it enables a quicktest of a supercapacitor’s capacitance. The proposedmass-balancingmethod is a simple andrelatively general method to optimize the capacitance of EDLCs. Althoughmass balancingmaynot achieve the highest possible capacitance, it can easily improve the resulting value. Exploit-ing a material’s full potential requires a time-consuming analysis of the components. Becausethis is not the aim of this study, we did not use any advanced electrochemical characterizationmethods.
Materials and Methods
Materials
The electrodematerials, graphite (ABG 2025, SO# 5-42-25, from Superior Graphite) and acti-vated carbon (Pulsorb 208CP Powder, from Chemviron Carbon), were used as received. Sulfu-ric acid (analytical grade, from VWR), sodium hydroxide (analytical grade, from VWR),potassium hydroxide (from Fluka) and sodium sulfate (analytical grade, fromMerck) werediluted with deionizedwater to produce electrolytes with a concentration of 1. Greaseproofpaper (untreated, grammage: 45 g/m2, from Nordic Paper) was employed as a separator andwas used as received.
Preparation of TEMPO-oxidized cellulose nanofibers
The process of the TEMPO-mediated oxidization of the CNFs followed the method describedby Saito et al. [18]. We used 100 fully bleached softwoodKraft pulp. The pulp was diluted in 10deionizedwater to 2.5% consistency. Then, 2 sodium bromide (NaBr, fromMerck Millipore)per g dry pulp and 0.2 mmol/g TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy, from Sigma-Aldrich) was added to the pulp suspension. The suspension was mixed, and 10 sodiumhypo-chlorite (NaClO, 14%, from VWR) per g dry pulp was added during stirring. To avoid largepH variations, small amounts of NaClO were added. The pH of the pulp suspension wasadjusted to 9.5 by adding 1 M NaOH. The addition of NaClO and the subsequent pH adjust-ment were performed for 2 to 3 hours. Afterwards, the pulp was dewatered and thoroughlywashed. The washed pulp was diluted to 1% consistency with deionizedwater. Approximately700 ml of the pulp suspension was dispersedwith the IKA T 25 digital Ultra-Turrax disperser(rotor: S25N-25F) at 15000 rpm for 30 minutes for proper mixing. This rotational speed corre-sponds to a shear velocity (circumferential speed) of 14.14m/s.
Calculation of ion size ratio and electrode mass ratio
The following assumptions were made to develop a simplifiedmodel of the optimal electrodemass ratio. The objective of this study is to determine the optimal electrodemass ratio in orderto increase the EDLC capacitance. However, we do not seek to fine-tune the mass ratio or thecapacitance. Thus, a simplifiedmodel is sufficient for our purpose.
The ion size ratio and the theoretical electrodemass ratio were calculated from the size ofthe hydrated electrolyte ions. A table showing the radii of various hydrated ions can be foundin [19]. Table 1 is an extract of this table and lists the ions tested in this study.
For an electrolyte molecule with the chemical formula Ax By, the ion size ratio is calculatedaccording to
ion size ratio ¼x � rAy � rB
; ð1Þ
where x is the number of ion A in the molecule, y is the number of ion B in the molecule, rA is
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 4 / 12
the radius of the hydrated ion A, and rB is the radius of the hydrated ion B. The theoretical elec-trodemass ratio is the inverse of the ion size ratio:
theoretical electrode mass ratio ¼y � rBx � rA
: ð2Þ
The electrodemass ratio of the tested EDLCs was calculated by
electrode mass ratio ¼mþm�
; ð3Þ
wherem+ is the mass of active material in the positive electrode, andm− is the mass of activematerial in the negative electrode.
The largest increase in capacitance can be expected for electrolytes with a significant differ-ence in ion size, i. e., a large anion and a much smaller cation (or vice versa). Moreover, onehas to consider the valence of the ions because it determines the number of anions and cationsthat can form the electric double layer at the electrode-electrolyte interfaces.We expect thelargest enhancement for sodium sulfate, which has a theoretical electrodemass ratio of 0.53,when considering the valence of the ions.
In general, electrolytes with small ions achieve larger capacitances than electrolytes com-posed of large ions at a given electrode size and weight. This result can be explained by a highercharge density at the electrode surface. Thus, the size of the electrolyte ions is crucial to pro-duce EDLCs with large specific capacitances. Moreover, the shape of the hydrated ions shouldbe considered because it influences the formation of the electric double layer.
Preparation of electrodes
All of the electrodeswere composed of the same composite. To prepare the electrodes, equalamounts of graphite and activated carbon were mixed with an additional 10% TEMPO-oxi-dized CNFs (10% of the total mass of graphite and activated carbon). Samples with a totalweight of active material between 0.2 g and 0.7 g were prepared. Approximately 40 of deionizedwater was added to each mixture. The suspensions were dispersed for 10 minutes at 12000 rpmusing the Ultra-Turrax disperser (rotor: S25N-10G). This rotational speed corresponds to ashear velocity (circumferential speed) of 4.71 m/s. Subsequently, the dispersions were filtratedon Millipore Durapore Membrane Filters (filter type: 0.22 GV, diameter: 90 mm) using a vac-uum filtration funnel. Films with coating weights between 44 g/m2 and 154 g/m2 wereobtained. All filters were dried at room temperature and cut into 3 cm × 3 squares.
Characterization of electrode materials
The specific surface areas (SSAs) of the raw materials, graphite, activated carbon, and the com-posite were analyzed. Measurements according to the Brunauer-Emmett-Teller (BET) theorywere performed using a Micromeritics Gemini 2370 BET instrument. Furthermore, digital
Table 1. Radii r of hydrated ions [19].
ion r/nm
cations H+ 0.282
Na+ 0.358
K+ 0.331
anions OH− 0.300
SO42− 0.379
doi:10.1371/journal.pone.0163146.t001
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 5 / 12
images of the composite surface were obtained using a field emission scanning electronmicro-scope (FESEM), ZeissMerlin FESEM. Secondary electron images (SEIs) were generated using a5 kV accelerating voltage and an in-lens detector.
Assembly and testing of electric double-layer capacitors
The EDLCs were assembled by stacking electrodes and a separator in a test cell. The test cell isdescribed in detail elsewhere [16]. Greaseproof paper was used as the separator between theelectrodes. The components were wetted with electrolyte prior to assembling the test cell.EDLCs with different electrodemass ratios were prepared, and galvanostatic cycling was per-formed by using a LabVIEW-based PXI system. The EDLCs were cycled for 24 hours between0 and 1V with a charge and discharge current of 8. The resulting discharge times and thus theamount of cycles per measurement depend on the EDLC capacitance and the applied current;see Eq (4). We selected a low current to allow comparison of devices with a high capacitance toEDLCs exhibiting a low capacitance. This low current resulted in discharge times between 150and 665s giving 65 to 288 cycles per 24-hour measurement.We chose a test duration of 24hours to evaluate the conditioning, efficiencyand cycling stability of the EDLCs. To verify theresults and eliminate the influence of the electrode thickness, five EDLCs with the optimal elec-trodemass ratio but varying electrodemass loadings were tested for each electrolyte. No signif-icant variations were observed.The mean value of the specific capacitances was calculated.Furthermore, measurements of the EDLCs with specific capacitances close to the maximumwere repeated by testing new EDLCs with the same electrodemass ratio. No significant varia-tions were observed for the replicates.
Capacitance, specific capacitance and efficiency. Galvanostatic cycling was performed toevaluate the capacitance of the EDLCs. The capacitance C was calculated from the dischargecurves according to
C ¼ I �dtdV
; ð4Þ
where I is the discharge current, dt is the discharge time, and dV is the cell potential difference.The specific capacitance Csp was calculated by
Csp ¼ 4 �Cm
; ð5Þ
where C is the capacitance of the EDLC andm is the mass of the active material in both elec-trodes. The factor of 4 adjusts the cell capacitance C to the mass and capacitance of one elec-trode, considering that each EDLC consists of two capacitors, one on each electrode. In thecase of our electrodes, only graphite and activated carbon were counted as active materials.CNFs were not considered as an active electrodematerial.
The efficiencyof the EDLCs was calculated from the galvanostatic cycling results by divid-ing the discharge time by the charge time.
Results
Characterization of electrode materials
Specific surface area. The specific surface area of the electrode components graphite, acti-vated carbon, and the produced composite was obtained by BET measurements. The results ofthese measurements are displayed in Table 2. As expected, the activated carbon exhibited ahigh specific surface area, whereas graphite had a low specific surface area. The composite,
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 6 / 12
which contains equal amounts of activated carbon and graphite and an additional 10% CNFs,exhibited an intermediate specific surface area of 418 m2/g.
Surface structure. The surface of the composite film was investigated using an FESEM.Fig 1 shows three images of different areas of the composite surface. Fig 1a and 1b show the lay-ered graphite flakes, the spherical activated carbon and the network of CNFs. Fig 1c shows aclose-up image of the nanofiber network. One can observe that the CNFs form a network thatconnects the graphite flakes and the activated carbon particles. This observation confirms thesuitability of CNFs as a binder for graphite and activated carbon electrodes.
Optimal electrode mass ratio
Fig 2 shows the results of the capacitance measurements. Depending on the electrolyte used,we could detect an influence of the electrodemass ratio on the specific capacitance of the
Fig 1. Structure of the composite surface taken with a field emission scanning electron microscope.
The layered structure of the graphite flakes, the spherical shape of the activated carbon, and the cellulose
nanofiber network are shown. a) magnification of 25000 ×; b) and c) magnification of 100000 ×.
doi:10.1371/journal.pone.0163146.g001
Fig 2. Influence of electrode mass ratio on specific capacitance of supercapacitors with different
aqueous electrolytes.
doi:10.1371/journal.pone.0163146.g002
Table 2. Specific surface area (SSA) of electrode materials.
material SSA/m2 g−1
activated carbon 903
graphite 19.8
composite 418
doi:10.1371/journal.pone.0163146.t002
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 7 / 12
EDLCs. For each electrolyte, an optimal electrodemass ratio could be extracted from the max-ima of the displayed curves; see Table 3. If more or less active material was used in either thepositive or the negative electrode, only lower specific capacitances could be achieved. In addi-tion, we detected large differences in the specific capacitance when comparing the differentelectrolytes. EDLCs with sulfuric acid as the electrolyte exhibited much higher capacitancesthan devices operating with one of the other electrolytes; see Fig 2 and Table 3. This findingcan be explained by the small size and highmobility of the hydrogen cations. The advantage ofsmall ions is that they can easily enter and fill small pores, which results in a higher capacitance.Large ions might not be able to enter narrow pores, leading to lower capacitances. Sodium sul-fate exhibited the lowest specific capacitance of all the tested electrolytes. Sodiumhydroxideand potassium hydroxide producedmoderate specific capacitances with similar values formost tested electrodemass ratios. This result was expected because of the similar radii ofhydrated potassium and sodium ions.
Sulfuric acid and sodium sulfate exhibited the highest specific capacitance at an electrodemass ratio of 0.75. For sodium hydroxide, the highest specific capacitance was detected at anelectrodemass ratio of 0.86. Potassium hydroxide, however, exhibited its largest specific capac-itance with supercapacitors having equal electrodemasses. An optimal electrodemass ratioequal to 1 or slightly below 1 was expected because potassium ions and hydroxide ions havesimilar ion radii. The largest increase in specific capacitance obtained by mass balancing wasdetected for sodium hydroxide. An increase of 8.6% was achieved by shifting the electrodemass ratio from 1 to 0.86; see Table 3. In the case of sodium sulfate and sulfuric acid, increasesof 3.7% and 1.5% were obtained. The four curves reveal different curve progressions. The curvecorresponding to sodium sulfate, in particular, differs from the other plots. For electrodemassratios larger than the optimal mass ratio, the specific capacitance declinedmore slowly than forthe other electrolytes. Potassium hydroxide, sodium hydroxide and sulfuric acid produced sim-ilar curves.
To eliminate the influence of the electrode thickness or the amount of available active mate-rial on the optimal electrodemass ratio, five EDLCs with the optimal electrodemass ratio butvarying electrodemass loadings were tested for each electrolyte. The specific capacitance of thecorrelating EDLCs deviated only slightly, indicating that the influencewas negligible.
In Table 3, the theoretical optimal electrodemass ratios are compared with the measuredvalues. The experimental results were approximately consistent with the theoretical optimalelectrodemass ratio calculated from the ion size ratio. However, slight deviations due to mea-surement inaccuracywere found. In the case of sodium sulfate, the theoretical optimal elec-trodemass ratio differed from the measured ratio to a greater extent.
Efficiency
In addition to the capacitance, we also measured the efficiencyof the EDLCs from the chargeand discharge curves. Fig 3 shows the efficiencyof supercapacitors operating with different
Table 3. Theoretical optimal electrode mass ratio emrth, measured optimal electrode mass ratio
emrm, highest specific capacitance Csp, and specific capacitance increase Cincr of different aqueous
electrolytes.
electrolyte emrth emrm Csp/F g−1 Cincr/%
Na2SO4 0.53 0.75 49.3 3.7
H2SO4 0.67 0.75 89.7 1.5
NaOH 0.84 0.86 66.8 8.6
KOH 0.91 1 67.6 0
doi:10.1371/journal.pone.0163146.t003
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 8 / 12
electrolytes at their optimal electrodemass ratio. All tested EDLCs exhibited a good cycling sta-bility. EDLCs with sulfuric acid or sodium sulfate reached efficiencies of approximately 99%;see Fig 3a. Devices with sodium hydroxide or potassium hydroxide produced slightly lowerefficiencies of 97%. After only five charge and discharge cycles, efficiencies of 97–98% forsodium sulfate and sulfuric acid and 94–95% for sodiumhydroxide and potassium hydroxidewere reached; see Fig 3b. The fast conditioning of the EDLCs can also be observed in Fig 4aand 4b. Fig 4 displays the first and last cycles of a typical 24-hour galvanostatic cycling test. Acomparison of the two graphs revealed that the first cycle differed from the following cycles.The curvewas bent during charging and discharging. During continuation of the measure-ments, the curves approached a more linear curve progression. In particular, the fast condition-ing of the EDLCs was noteworthy. However, the efficiencywas not influenced by the electrodemass ratio.
Discussion
Althoughminimal enhancement in specific capacitance was observed for some electrolytes,mass balancing can be a simple approach for increasing the capacitance of EDLCs. For devicesusing suitable electrolytes, such as sulfuric acid, mass balancing can significantly improveEDLC performance. The mass-balancing approach can also be applied to other electrode-elec-trolyte combinations and can be implemented in large-scale production processes without anydifficulty. Mass balancing further leads to a reduction in unusedmaterial and thus a reductionof the weight and volume of EDLCs.
Although one can observe large differences in the specific capacitance between the testedelectrolytes, caution should be used when choosing an electrolyte. Sulfuric acid provided the
Fig 3. Efficiency of electric double-layer capacitors operating in different electrolytes at the
corresponding optimal electrode mass ratio. a) efficiency for each cycle of a 24-hour measurement and
b) development of the efficiency during the first 10 cycles of the same measurements.
doi:10.1371/journal.pone.0163146.g003
Fig 4. Galvanostatic cycling profile of an electric double-layer capacitor. a) the first cycles and b) the
last cycles of a typical 24-hour measurement.
doi:10.1371/journal.pone.0163146.g004
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 9 / 12
highest specific capacitance in this study, although it might not be the best choice for all super-capacitors. Generally, one should consider the influence of the electrolyte on supercapacitormaterials, such as the electrodes, the separator and the passive components. The pHmight becrucial in terms of electrode stability, separator stability and corrosion. Neutral electrolytesmight be favorable in some supercapacitors. For our system, in which cellulose-basedmaterialswere used, neutral electrolytes, e. g. sodium sulfate, are beneficial.However, when acidic orbasic electrolytes can be employed, one should use an electrolyte that generates a higher specificcapacitance, such as sulfuric acid. Moreover, the performance of the electrolyte is influenced bythe electrodematerial and structure. Thus, recommendations can only be given for particularelectrode-electrolyte systems. However, the mass-balancing approach can be applied on allsupercapacitors, regardless of the electrode or electrolyte materials used in the system.
To further optimize an EDLC’s performance, one should measure the surface area that isaccessible for the anions or cations rather than the electrodemass or the pore size distribution.Narrow pores might only be attainable for small ions. In this case, small ions will have a largerusable surface area than large ions. Having an ion pair with large differences in ion size mightresult in two different accessible surface areas. Thus, the accessible surface area should be con-sidered to obtain accurate calculations. However, it is not simple to measure the usable surfacearea for particular ions. Therefore, using the electrodemass to obtain a well-balanced cell ismore practical.
Furthermore, one should choose a higher charge and discharge current when performinggalvanostatic cycling to obtain faster discharge times that reflect the intended application inKERSs.
Further studies should be conducted to determine the influence of other electrolytes on theoptimal electrodemass ratio. The mixture of two or more compatible electrolytes might also beinteresting. In addition to aqueous electrolytes, even organic electrolytes and ionic liquidscould be investigated.
Conclusions
In this study, we showed that the capacitance of EDLCs can be increased if the electrodemassratio is adjusted to the ion size ratio. This optimization is favorable for improving the perfor-mance of EDLCs, reducing the amount of unused electrodematerial and thus decreasing thecell weight, volume and cost. The highest increase in specific capacitance could be achieved forEDLCs using sodium hydroxide as the electrolyte. An increase of 8.6% was obtained by shiftingthe electrodemass ratio from 1 to 0.86. The highest specific capacitance was obtained withEDLCs using sulfuric acid as the electrolyte. By changing from a symmetric electrode configu-ration to the optimal electrodemass ratio of 0.75, only a small increase in specific capacitanceof 1.5% could be obtained for sulfuric acid.
Acknowledgments
The authors acknowledge the facilities and technical assistance of the Umeå Core Facility forElectronMicroscopy (UCEM), Umeå University.
Author Contributions
Conceptualization:BA SF.
Data curation: BA.
Formal analysis:BA ACE NB SF CD.
Mass Balancing as an Inexpensive and Simple Method to Increase the Capacitance of EDLCs
PLOS ONE | DOI:10.1371/journal.pone.0163146 September 22, 2016 10 / 12
Funding acquisition: SF HO.
Investigation: BA ACE NB CD.
Methodology:BA SF.
Project administration: SF HO.
Resources:BA ACE NB.
Supervision:SF HO.
Validation: BA SF.
Visualization: BA CD.
Writing – original draft: BA.
Writing – review& editing: BA ACE NB SF CDHO.
References1. Zhu Y, Murali S, Stoller M D, Ganesh K J, Cai W, Ferreira P J, et al. Carbon-based supercapacitors
produced by activation of graphene. Science. 2011; 332:1537. doi: 10.1126/science.1200770 PMID:
21566159
2. Chmiola J, Yushin G, Gogotsi Y, Portet C, Simon P, Taberna P L. Anomalous increase in carbon
capacitance at pore sizes less than 1 nanometer. Science. 2006; 313:1760. doi: 10.1126/science.
1132195 PMID: 16917025
3. Jeong H M, Lee J W, Shin W H, Choi Y J, Shin H J, Kang J K, et al. Nitrogen-doped graphene for high-
performance ultracapacitors and the importance of nitrogen-doped sites at basal planes. Nano Lett.
